Research: Educational and Psychological Aspects Diabetes-specific self-esteem, self-care and glycaemic control among adolescents with Type 1 diabetes

Abstract

Aims

To explore the relationships between diabetes-specific self-esteem, self-care and glycaemic control among diverse adolescents with Type 1 diabetes.

Methods

Adolescents (aged 13–21 years) diagnosed with Type 1 diabetes for at least one year, receiving care at an urban medical centre, completed a self-report battery including demographic information, the Diabetes-Specific Self-Esteem Scale and Self-Care Inventory. Glycaemic control (HbA_1c) was obtained from medical records at recruitment and one year later. Bivariate correlation and multiple linear regression assessed relationships between the Diabetes-Specific Self-Esteem scale, Self-Care Inventory and HbA_1c at baseline and one year.

Results

Participants included 85 adolescents (15.9 ± 2.1 years; 53% women; 47% Hispanic/Latino). Diabetes-specific self-esteem scores did not differ based on patient characteristics but were significantly correlated with baseline self-care (r = 0.59, P < 0.001) and HbA_1c at baseline (r = −0.51, P < 0.001) and one year later (r = −0.48, P < 0.001). Diabetes-specific self-esteem remained a significant correlate of baseline (β= −0.41, P < 0.001) and follow-up HbA_1c (β= −0.37, P = 0.008) when adjusting for covariates and self-care. Diabetes-specific self-esteem was not significantly associated with change in HbA_1c.

Conclusions

Results suggest that diabetes-specific self-esteem is significantly associated with self-care and glycaemic control among diverse adolescents with Type 1 diabetes. Diabetes-specific self-esteem may be more closely related to HbA_1c than reports of the frequency of self-care behaviours, and could represent a useful tool for clinical and research applications.

Introduction

A crucial development during adolescence is self-concept formation and the emergence of an independent identity. Self-esteem, defined as a subjective evaluation of the self, is a central component of one’s self-concept. It is conceptualized as an attitude and judgement of one’s worth and serves as an appraisal of one’s value, either generally or within a specific area of life (e.g. academic, sport) [1]. Self-esteem also involves feelings or affective responses to the self; high or low self-esteem is determined based on an individual’s sense of whether their ‘actual self’ meets what the individual has aspired to be [2,3].

Although related to self-efficacy, or an individual’s belief about their ability to perform a particular behaviour, the literature supports self-esteem and self-efficacy as distinct constructs. Chen and colleagues [2] demonstrated that general self-efficacy was closely related to motivational states and achievement, while general self-esteem was more related to affective factors, including anxiety, negative affect, avoidance and emotional stability. While researchers examining self-efficacy in adolescents with Type 1 diabetes have found support for a clear relationship between greater diabetes-specific self-efficacy and better diabetes outcomes [4–6], minimal research has been undertaken to test the relationship between global or illness-specific self-esteem and diabetes outcomes.

Yet research has shown that self-esteem may play an important role in the management of chronic illnesses, with data suggesting self-esteem serves as a stronger predictor of life outcomes rather than as a consequence of them. Researchers found that self-esteem was predictive of physical health over time; however, physical health did not significantly contribute to self-esteem across the lifespan in either adults or adolescents [7,8]. The relevant literature provides similar support for predictive validity, such that a positive self-concept, which includes general self-esteem, was associated with better glycaemic control in adolescents with Type 1 diabetes [4,9]. Luyckx and Seiffge-Krenke [10] also found that adolescents with optimal control could be differentiated from those with deteriorating control based on their overall self-concept scores, with individuals with optimal control having the highest self-concept ratings. As such, it is possible that the acquisition of high self-esteem in the context of a chronic illness may serve as a buffer against factors that may pose a risk to disease outcomes.

Conclusions drawn from the literature assessing self-esteem and glycaemic control are significantly limited by the measurement of self-esteem. Self-esteem is predominantly assessed with global measures, despite evidence supporting a multidimensional framework for self-esteem [11] and a well-supported specificity-matching principle. This principle indicates that stronger relationships are observed when global predictor variables (e.g. global self-esteem) are matched to global outcomes (e.g. overall quality of life) and specific predictors (e.g. diabetes-specific self-esteem) are matched to specific outcomes (e.g. HbA_1c) [12]. Aligned with this expectation, Schneider and colleagues [3] created and validated the Diabetes-Specific Self-Esteem Scale (DSSE) to assess an individual’s sense of worth and subjective evaluation of themselves within the context of Type 1 diabetes. Diabetes-specific self-esteem was defined as ‘the degree of self-satisfaction or self-approval in relation to the medical and social management of diabetes in everyday life’ [3]. Results of this study indicated that the DSSE had acceptable concurrent and incremental validity in relation to global self-esteem. Factor analysis revealed a distinct diabetes domain, with diabetes-related items uniquely contributing to an adolescent’s self-esteem. These results demonstrated that young people with Type 1 diabetes have a distinct sense of self that is not fully captured by the domains presented in a measure of global self-esteem (e.g. peers, school). Additionally, higher scores on the DSSE were significantly and independently associated with better diabetes self-care and HbA_1c after accounting for the effects of global self-esteem.

While the initial validation of the DSSE provides preliminary evidence supporting the role of illness-specific self-esteem, there are significant limitations to our understanding of the role of diabetes-specific self-esteem in individuals with Type 1 diabetes. Research on global self-esteem indicates that self-esteem differs based on a number of important demographic factors. Self-esteem continues to change between middle adolescence and young adulthood (14–23 years), when men tend to show an increase and women a decrease in self-esteem [13]. Self-esteem trajectories also differ based on ethnicity, with African American adolescents reporting higher self-esteem compared with their Latino and Asian peers [14]. However, participants in the original DSSE validation study were aged 10–16 years and were primarily white Europeans, thus limiting the generalizability of the results.

Furthermore, age and race/ethnicity have been shown to significantly impact overall diabetes care. Adolescence has long been identified as a developmental stage during which glycaemic control and self-care worsen [6,15], but research has shown that ethnic-minority status significantly increases the probability of poor self-care and glycaemic control [16–18]. Increasing knowledge about the relationship between self-esteem and diabetes management in a high-risk group has the potential to further increase our understanding of the decline in self-care and glycaemic control, as well as to inform future research and interventions for providers who work with those adolescents who are struggling to maintain optimal control from a physical and emotional perspective.

The primary goal of the current study was to replicate and extend the initial validation results reported for the DSSE [3] in a diverse and understudied population of adolescents. We examined DSSE scores in relation to demographic, illness and treatment factors in a sample of primarily socioeconomically disadvantaged racial and ethnic minority youth. We hypothesized that greater diabetes-specific self-esteem would be significantly associated with better self-reported diabetes self-care and glycaemic control, assessed cross-sectionally and one year later. We explored race and ethnicity as a moderator of these relationships.

The secondary goal was to examine whether diabetes-specific self-esteem is independently associated with HbA_1c when diabetes self-care is controlled. While the DSSE was developed with content guided by a theoretical framework for domain-specific self-esteem [12], the item content overlaps considerably with the item content of typical diabetes self-care measures. The contribution of the DSSE as a correlate and predictor of glycaemic control, independent of diabetes self-care, has not been examined previously and would be important in strengthening the evidence for construct validity. We hypothesized that diabetes-specific self-esteem would be significantly associated with glycaemic control, independent of self-reported diabetes self-care.

Methods

Participants and procedures

All participants were part of a larger mixed-methods study including qualitative interviews and validated self-report measures. Eighty-nine adolescents were approached by study staff during their standard endocrinology clinic appointments at an urban academic medical centre. Of the 89 adolescents, 85 completed the study (2 declined to participate; 2 consented but not did not complete the measures). This medical centre primarily serves residents living in one of the nation’s poorest urban counties; the poverty rate is 31.0% (compared with 19.4% city-wide) with a median income of $32 450. In order to be eligible, participants had to be aged 13–21 years, living with diabetes for at least one year, able to speak and read in English, and be without any psychiatric or medical co-morbidity that could affect their assent/consent or contribution to the study.

Data collection

Written consent from the parent and assent from the child were obtained in person for participants aged <18 years before completion of self-report measures. Informed consent was obtained for participants aged ≥18 years. The Albert Einstein College of Medicine Institutional Review Board approved all study procedures.

Measures

Demographic information

Participants provided demographic information as part of the study questionnaires. Variables assessed in the current study included: age, gender, duration of illness, race/ethnicity (black or African American; Hispanic or Latino; white European; native American; Asian; other) and insulin delivery method (pump or injections).

DSSE – youth report version

The DSSE [3] contains 10 items to measure and assess an adolescent’s subjective self-reported ratings of their performance of diabetes self-care tasks, for instance, ‘I am as good as I want to be at taking care of my illness’ and ‘I am satisfied with how often I test my blood sugars’. Items are scored on a four-point Likert scale ranging from 1 (strongly disagree) to 4 (strongly agree), with higher scores indicating higher diabetes-specific self-esteem. The DSSE was originally validated on a sample of 87 youths with Type 1 diabetes and their parents, and demonstrated high internal consistency (youths α= 0.87), with significant agreement between youth and parent reports (intraclass correlation coefficient = 0.41, P = 0.01). Scores in the current sample were normally distributed (skewness_z = 0.19, kurtosis_z = −1.26) and demonstrated high internal consistency (α= 0.86).

Self-Care Inventory

The Self-Care Inventory (revised version) [19] is a widely used 15-item assessment of self-care behaviours based on treatment recommendations of the American Diabetes Association. Participants are asked to rate the frequency with which they perform diabetes-specific behaviours (e.g. ‘check blood glucose with monitor’, ‘take the correct dose of insulin’) on a five-point Likert scale ranging from 1 (never) to 5 (always). Cronbach’s alpha was high in the current sample (α= 0.87).

HbA_1c

Consent was obtained from each participant or parent during the time of recruitment to access the HbA_1c results. Follow-up HbA_1c was obtained 9–15 months post-baseline.

Statistical analysis

A power analysis was conducted using G-Power to determine sufficient sample size for linear regression with up to eight predictor variables. Assuming a medium effect size (f² = 0.15), the desired sample size was 89. Effect-size expectations were based on the effect size for the relationships among self-care, psychosocial factors and HbA_1c in youth with Type 1 diabetes [20,21]. The distribution of variables was examined and non-normally distributed variables were transformed and used in all subsequent analyses. Means, SD and sample size were used to compare DSSE scores between the current study and Schneider et al. [3]. Categorical differences on DSSE scores and HbA_1c were examined with independent sample t-tests. In order to examine potential differences by race and ethnicity, participant responses were coded as Hispanic/Latino (yes/no), non-Hispanic black or African American (yes/no) and non-Hispanic white European (yes/no). Associations between continuous variables were evaluated with Pearson’s correlation. Difference scores between baseline and follow-up HbA_1c were calculated for each participant, with positive change scores indicating an increase in HbA_1c. To examine whether race/ethnicity moderated the relationship between DSSE and HbA_1c, three moderation analyses were run using PROCESS [22].

To examine whether diabetes-specific self-esteem predicted glycaemic control, hierarchical multiple regression analyses were conducted with baseline and follow-up HbA_1c. To determine if the DSSE predicted change in HbA_1c, a third regression analysis was conducted with follow-up HbA_1c as the outcome and baseline HbA_1c as a covariate. For each model, predictors included self-care, self-esteem and covariates. Potential covariates identified in the literature or those that showed a significant association were included. All analyses were conducted in SPSS version 22 (IBM Corp; Armonk, NY, USA). Statistical significance was determined at the 0.05 level. No allowance was made for multiplicity of statistical testing.

Results

Preliminary analyses

Eighty-five adolescents aged 13–21 years participated in the current study. Demographic and disease characteristics are provided in Table 1 (data are expressed as mean ± SD). Comparison of DSSE scores for the current sample against those reported by Schneider and colleagues [3] shows that the current sample’s scores were significantly lower (2.5 ± 0.6) than the scores reported by Schneider and colleagues (2.8 ± 0.6, P = 0.0002). Self-esteem did not differ significantly by gender, although there was a trend (P = 0.07) for women to report lower self-esteem (2.4 ± 0.6) than men (2.6 ± 0.5). DSSE scores were not significantly different based on the insulin administration method, although there was a trend (P = 0.051) indicating that those using a pump had higher DSSE scores (2.6 ± 0.6) compared with those using injections (2.3 ± 0.6). DSSE scores were not significantly associated with age or duration of illness. Self-esteem was significantly correlated with self-care and HbA_1c at baseline and after one year, but showed no association with change in HbA_1c (Table 2). We further examined potential differences by race/ethnicity in diabetes-specific self-esteem, comparing those who identify as Hispanic/Latino or not, non-Hispanic black/African American or not, and non-Hispanic white European or not. There were no significant differences in DSSE scores based on race/ethnicity (P-values > 0.23).

Table 1.

Sample characteristics (n = 85)

Variable	N (%)	Mean (SD)	Range
Age (years)		15.9 (2.1)	13–21
Gender
Women	45 (53)
Men	40 (47)
Race
African American	24 (28)
White European	17 (20)
Asian	3 (4)
Hispanic/Latino	40 (47)
Other	1 (1)
Duration of illness (years)		6.7 (3.7)	1–15
Insulin method
Pump	47 (55)
HbA1c
Baseline		87 (25) mmol/mol10.1 (2.3)%	48–147 mmol/mol6.5–15.6%
Follow-up (n = 70)		91 (23) mmol/mol10.5 (2.1)%	48–158 mmol/mol6.5–16.6%
Diabetes- specific self-esteem		2.5 (0.6)	1.3–3.8
SCI		58.6 (16.4)	26.7–91.7

Notes. Self-Care Inventory (SCI) scores are converted to a 0–100 point scale, with higher scores indicating better self-care. Diabetes-specific self-esteem (DSSE) scores range from 1–4 with higher scores indicating higher diabetes-specific self-esteem.

Table 2.

Correlations among self-esteem, self-care and glycaemic control

	1	2	3	4	5	6	7
1. Self-esteem		0.59	0.12	0.05	−0.51	−0.48	0.12
2. Self-care			−0.13	0.10	−0.40	−0.33	0.15
3. Age (years)				0.03	−0.08	−0.01	−0.09
4. Duration of illness					−0.13	0.04	0.28
5. HbA1c baseline						0.73	−0.48
6. HbA1c follow-up							0.27
7. Change in HbA1c

Notes. bold items indicate significance at the P < 0.05 level.

Those using an insulin pump had significantly lower HbA_1c than those using injections at baseline [80 ± 25 mmol/mol (9.5 ± 2.1%) vs. 93 ± 23 mmol/mol (10.7 ± 2.5%), P = 0.02] and follow-up [(84 ± 19 mmol/mol (9.8 ± 1.8%) vs. 100 ± 25 mmol/mol (11.3 ± 2.3%) P = 0.003], respectively. Longer duration of illness was statistically significantly associated with an increase in HbA_1c from baseline to follow-up (r = 0.28, P = 0.02). There were no significant differences based on race/ethnicity for baseline HbA_1c (P-values > 0.09) or follow-up HbA_1c (P-values > 0.15).

For exploratory analyses, we ran three moderation models using PROCESS to assess whether identifying as Hispanic/Latino (yes/no), non-Hispanic black/African American (yes/no) or non-Hispanic white European (yes/no) moderated the relationship between DSSE and HbA_1c. None of the models showed a significant moderation effect (P-values > 0.38; data not shown).

Regression analyses

When entered into separate regression models, self-care and diabetes self-esteem were independent correlates of HbA_1c at baseline and follow-up (Figures 1 and 2), even in the presence of covariates. When entered together, self-esteem remained significantly associated with baseline HbA_1c (β= −0.41, P = 0.001) and follow-up HbA_1c (β= −0.37, P = 0.008). Self-care was not significantly associated with HbA_1c in either model. Neither self-care (P = 0.87) nor self-esteem (P = 0.42) was significantly associated with change in HbA_1c (Table 3).

FIGURE 1.

Scatterplot of Diabetes-Specific Self-Esteem scale (DSSE) and baseline HbA_1c.

FIGURE 2.

Scatterplot of Diabetes-Specific Self-Esteem scale (DSSE) and follow-up HbA_1c.

Table 3.

Summary of hierarchical regressions predicting baseline HbA_1c, follow-up HbA_1c and follow-up HbA_1c controlling for baseline HbA_1c

	b	SE(b)	β	R2	∆R2
Baseline HbA1c
Step 1				0.11	0.11
Duration of illness	−0.01	0.01	−0.11
Insulin administration	−0.14	0.08	−0.19
Race/ethnicity	−0.11	0.10	−0.12
Age	−0.01	0.02	−0.07
Gender	0.09	0.08	0.13
Step 2				0.34	0.23
Duration of illness	−0.01	0.01	−0.07
Insulin administration	−0.05	0.07	−0.07
Race/ethnicity	−0.17	0.09	−0.19
Age	−0.004	0.02	−0.02
Gender	0.04	0.07	0.06
Self-care	−0.003	0.003	–0.13
Self-esteem	−0.25	0.08	−0.41*
Follow-up HbA1c
Step 1				0.18	0.18
Duration of illness	0.03	0.07	0.05
Insulin administration	−1.11	0.54	−0.26*
Race/ethnicity	−0.44	0.67	−0.08
Age	−0.09	0.14	−0.07
Gender	1.04	0.53	0.24
Step 2				0.33	0.15
Duration of illness	0.04	0.06	0.07
Insulin administration	−0.56	0.52	−0.13
Race/ethnicity	−0.84	0.62	−0.15
Age	−0.05	0.13	−0.04
Gender	0.89	0.48	0.21
Self-care	−0.01	0.02	−0.07
Self-esteem	−1.31	0.47	−0.37*
Change in HbA1c
Step 1				0.60	0.60
Duration of illness	0.09	0.05	0.15
Insulin administration	−0.14	0.40	−0.03
Race/ethnicity	−0.30	0.47	−0.06
Age	−0.09	0.10	−0.07
Gender	0.93	0.37	0.22*
Baseline HbA1c	4.21	0.51	0.71*
Step 2				0.61	0.01
Duration of illness	0.08	0.05	0.15
Insulin administration	−0.07	0.41	−0.02
Race/ethnicity	−0.41	0.48	−0.08
Age	−0.08	0.10	−0.07
Gender	0.90	0.37	0.21*
Baseline HbA1c	3.90	0.59	0.65*
Self-care	−0.002	0.01	−0.02
Self-esteem	−0.32	0.39	−0.09

^*items significant at the P < 0.05 level.

Discussion

The current study builds on the work of Schneider and colleagues [3] by further investigating the role of diabetes-specific self-esteem in a racially and ethnically diverse sample of adolescents with Type 1 diabetes. Results indicated that the DSSE retained good internal consistency and was statistically significantly associated with glycaemic control at baseline and after one year, even when accounting for self-reported diabetes self-care. In line with the multidimensional theory of self-esteem and the specificity-matching principle, the current study uniquely employed the use of a domain-specific measure of self-esteem.

The finding that a measure of diabetes-specific self-esteem showed cross-sectional and longitudinal associations with glycaemic control, independent of disease factors and self-care, is novel. These results suggest that those aspects unique to the DSSE, the assessment of one’s own emotional and cognitive impressions of self-care, may be more important to diabetes outcomes over time than assessments of the frequency of behaviours. A growing amount of literature supports that an individual’s perception of their overall health is consistently related to morbidity and mortality in individuals with and without diabetes [23]. Additional patient-reported outcomes, including diabetes distress [24], have also been strongly associated with diabetes outcomes, highlighting the important role of an individual’s lived experience with a chronic illness. It is likely that measures assessing quality of care allow for the inclusion of underlying factors that impact glycaemic control, which are often missed when measures focus solely on the frequency of behaviours.

While the longitudinal association suggests that self-esteem may contribute to glycaemic control, it is also important to note that feelings of self-worth are subject to context-specific appraisals [11]. Although the majority of adolescents are often unaware of their HbA_1c target [25], it is also possible that improved HbA_1c or positive feedback about self-care from medical providers may contribute to improvements in diabetes-specific self-esteem over time. Research has also indicated that self-esteem and outcome variables may engage in reciprocal relationships, as seen in a number of areas, including exercise behaviours and academic performance [11,26].

While the DSSE did not significantly predict change in HbA_1c, baseline HbA_1c was already significantly elevated, with HbA_1c values of 75 mmol/mol (9.5%) occurring in 57.8% of participants. HbA_1c was also relatively stable in our sample, with baseline and follow-up HbA_1c scores sharing 53% of their variance. Finally, follow-up HbA_1c was collected 9–15 months post-baseline, creating a small window for change. It is likely that these factors limited the potential for variation, impacting our ability to detect if diabetes-specific self-esteem was related to change in HbA_1c. Future studies should expand the time frame in which HbA_1c is collected and examine whether interventions aiming to improve disease-specific self-esteem also lead to improved glycaemic control.

Additionally, we did not observe any associations between the DSSE and demographic factors such as age, gender and race/ethnicity, which was inconsistent with the DSSE validation study and the wider general self-esteem research. One possible explanation for this difference may be the average age of each study sample’s participants; the current study included adolescents aged 13–21 years, thus capturing a different developmental period than the original study [3]. The literature on general self-esteem also shows inconsistent results regarding the effect of age on self-esteem. Some results show an increase in self-esteem during adolescence [7] while other data demonstrate a sharp decrease throughout adolescence [27]. Further examination of diabetes-specific self-esteem is needed to determine whether or not a pattern exists.

Comparison of DSSE scores for the current sample with those from Schneider and colleagues [3] shows that the current sample’s scores were, on average, significantly lower. The current study’s sample was more ethnically diverse and included more participants using insulin pumps than the previously reported sample; however, the duration of illness and percentage of female participants were both similar. Thus it is unclear which of these factors might explain the lower levels of diabetes-specific self-esteem, as we did not find any relationships among these demographic factors and the DSSE levels in our sample. It may be important to investigate interactions among these factors in predicting DSSE, as research shows that interactions between race, ethnicity and age are a stronger determinant of self-esteem than any one of these factors on its own [28].

Several limitations to the current study should be considered. Alhough we were able to obtain HbA_1c levels one-year post-baseline, the DSSE and Self-Care Inventory scores were only obtained at baseline, limiting our ability to examine how changes in these factors could have predicted changes in glycaemic control or the direction of these effects. Additionally, factors such as age, race/ethnicity and duration of illness were not significantly associated with self-care or glycaemic control, contrary to previous research [6,17,18]. While the homogeneity and small sample size may have made differences among groups or of moderating effects difficult to detect, the role of race/ethnicity should continue to be examined. However, it is also likely that these participants represent a unique sample of adolescents. Average HbA_1c was significantly higher than the HbA_1c of a national sample of over 16, 000 adolescents in the USA [15]. More recent research assessing trajectories of glycaemic control found that parents’ social status, and also distress, differentiated individuals with elevated HbA_1c from those with optimal glycaemic control, while gender, race/ethnicity and duration of illness did not [29]. Thus it may be useful for future studies to explore additional demographic factors (i.e. income, insurance) and their relationships with diabetes-specific self-esteem and diabetes outcomes.

Despite these limitations, the results of the current study provide data that support further exploration of the DSSE and diabetes-specific self-esteem more broadly. As noted by Schneider and colleagues [3], the existence of disease-specific self-esteem suggests the possibility of a shift in the perception of how diabetes impacts self-esteem, moving from negative self-esteem as a result of diabetes to an attainment of positive self-esteem as adolescents adapt to their individual environments in ways that include their diabetes care. This new perspective is reminiscent of research focusing on resilience in paediatric chronic illness that aims to identify factors and processes which help individuals successfully overcome the demands of a chronic illness. Resilience can be demonstrated in one or more domains, including disease management. The diabetes risk and resilience model [30] defines resilience as ‘a class of outcomes that are more positive than would be expected for an individual exposed to serious adversity’ (p. 739). Prior interventions developed to improve self-esteem have shown simultaneous improvement in body image [31] as well as behavioural, academic and emotional functioning [32]. However, further research is needed to identify both the direct and indirect links (e.g. behavioural and emotional pathways) through which self-esteem may impact upon glycaemic control.

This study highlights the utility of a disease-specific measure of self-esteem and emphasizes the importance of understanding an individual’s perception of their disease management. A strong relationship with disease outcomes may otherwise be missed if the frequency of behaviours is the primary focus. Given that adolescence is a challenging period for diabetes care, the inclusion of a tool that evaluates self-confidence in performing diabetes tasks, as well as identifying a modifiable variable that may improve outcomes, is crucial. Thus, future research should focus on identifying additional demographic and psychosocial variables associated with diabetes-specific self-esteem, as well as continuing to explore its relationships with diabetes outcomes and self-care.

What’s new?

• Prior research supports a diabetes-specific domain of self-esteem in adolescents with Type 1 diabetes. Our study evaluated the relationship between diabetes-specific self-esteem and diabetes outcomes in a sample of predominantly ethnic minority adolescents with Type 1 diabetes.

• We identified statistically significant relationships between diabetes-specific self-esteem and glycaemic control cross-sectionally and longitudinally. These relationships were independent of duration of illness, insulin administration and self-care.

• Findings highlight the importance of adolescents’ perceptions of their self-worth in the context of Type 1 diabetes and may be an important factor in identifying adolescents who are at risk of poor self-care and glycaemic control.